Results: 5 STL-V2s with XM-L2 de-domed

A few weeks ago I was asked to do some modifications to 1/2 a dozen SN6X-2s (otherwise known as the Skyray STL-V2 or Fandyfire STL-V6). These modifications basically melt down to swapping the XML for an XM-L2, increasing heat-sinking via a copper board and finally de-doming.

Due to the dimensions of the STL-V2 reflector I was not able to use copper so it was up to plan B:

I used 3/32” brass rod instead of copper due to time constrains and being unable to find anything suitable and this seams to have worked well.

The process was basically: remove 32mm MCPCB, drill 2 3/32” holes in the board, fill with 4-5mm of brass rod…. bash with hammer. Grind flush with the rest of the board (Dremel) and mount up a XM-L2 (U2 bin from Lightmall in this case) and de-dome. Finished.

Here are the results for the 5 I have done so far (still 1 to do):

Lux readings at 4.56m (x 20.8 for ~lux @ 1m)

Torch 1:
Power (tailcap): 2.10A x 8.2V
XML: 3000 lux
XM-L2: 3500 lux
XM-L2 de-domed : 7300 lux
XM-L2 de-domed with brass: 8300 lux ( so brass resulted in a 13% increase)
Final: ~172,000 lux @ 1m

Torch 2:
Power (tailcap):1.96A x 8.25V
XML: 2600
XM-L2: 3900 ~50% increase
XM-L2 de-domed: 8500
Final: ~177,000 lux @ 1m

Torch 3:
Power (tailcap):2.08A x 8.2V
XML: 2800
XM-L2: 3700 ~32% increase
XM-L2 de-domed: 7600
Final: ~158,000 lux @ 1m

Torch 4:
Power (tailcap):1.28A x 8.2V
XML: 2100
XM-L2: 3000 ~42% increase
XM-L2 de-domed: 7100
Final: ~157,000 lux @ 1m

Torch 5:
Power (tailcap):1.37A x 8.2V
XML: 2200
XM-L2: 2850 ~29% increase
XM-L2 de-domed: 6600
Final: ~137,000 lux @ 1m

Wow! Impressive. How is the tint after de doming?

Almost exactly the same as a XML U2 de-domed: warm white…

Good, as for the tint on my first dedome attempt, I am dissappont. Too much green, not to mention the nasty hot spot which might be a reflector positioning thing.
Anyway, you have some good results here so it is possible to do it right. I’ll have to mess with it some more before I completely give up.

Really nice results, better than double, consistently on the de-domes. What de-dome method did you use?

To de-dome I unscrewed the MCPCB… lifted it away from any metal contact and ran the torch on high until the board was too hot to hold (10-30 seconds dependent on torch).

Placing my finger as low down on the dome as possible on the opposite side of the bond wires I peel upwards… no need to rush this step as if the LED cools back down you can simply reheat the LED and continue. I have had very few failures with this method (just tried it on an XP-G2 and it worked fine)… of the 2 dozen or so I have done I think I have messed up 3: all of which were me rushing or not getting the temp right.

As far as the increase in lux: it seems to be almost exactly 110% every time.

Do you think the result with the brass (since brass is such a poor conductor), is because the die got that much hotter, lowering the Vf and creating a higher lux reading?

No… even though the brass is a poor conductor compared to copper or aluminium I think the gain comes from the removal of the dielectric between the LED pad and the base material. Solder itself has a worse thermal conductivity then brass so if anything that will be the main limiting factor.

Something I forgot to put in the original post: all readings were taken in the first 10 seconds of turn on so are not as representative of in the field performance… but I think this also masks some of the gain achieved by the brass rod.